Experimental browser for the Atmosphere
One really nice thing about all this neural net research is how much tools like speech-to-text (e.g. dictation) have improved. In a few minutes, you can download open source code running free models that can transcribe audio faster than real time (on a mid-range laptop) in better quality than e […]
Jun 18, 2024, 12:38 AM
{ "cid": "bafyreiast3mkafhlengeoq235t5zehzk6lurojpuwwekyak3rvbdlrd5tq", "uri": "at://did:plc:qnsm4tda5hprg3t623luuufu/app.bsky.feed.post/3kv67ohiyocp2", "value": { "$type": "app.bsky.feed.post", "bridgyOriginalText": "<p>One really nice thing about all this neural net research is how much tools like speech-to-text (e.g. dictation) have improved. In a few minutes, you can download open source code running free models that can transcribe audio faster than real time (on a mid-range laptop) in better quality than e.g. YouTube from 4 years ago.</p><p>I'm very excited about never needing to *listen* to a podcast ever again! ^-^</p>", "bridgyOriginalUrl": "https://mastodon.social/@fontenot/112634830916558280", "createdAt": "2024-06-18T00:38:46.000Z", "embed": { "$type": "app.bsky.embed.external", "external": { "$type": "app.bsky.embed.external#external", "description": "", "title": "Original post on mastodon.social", "uri": "https://mastodon.social/@fontenot/112634830916558280" } }, "text": "One really nice thing about all this neural net research is how much tools like speech-to-text (e.g. dictation) have improved. In a few minutes, you can download open source code running free models that can transcribe audio faster than real time (on a mid-range laptop) in better quality than e […]" } }